**First — extract repeat flags:** Follow the REPEAT PREAMBLE from `mastermind-repeat/SKILL.md`. Parse `$ARGUMENTS` for `--auto`, `--confirm`, `--project `, and remaining text = problem description. Load brain context for the `debug` domain (follow `mastermind-protocol/SKILL.md` Brain Load Procedure). If problem description is empty: ask "What issue are you debugging?" Default mode: **auto** (proceed immediately; the skill gates on root-cause investigation, not on user confirmation). --- Invoke `Skill("mastermind-debug")` passing: brain_context, problem_description, mode. After skill returns: follow `mastermind-protocol/SKILL.md` Brain Write Procedure for domain `debug`. **MANDATORY — invoke `Skill("mastermind-repeat")` now.** This is required regardless of how the skill above completed, regardless of whether you think the work is done, regardless of whether you plan to end your response. For `--repeat N`: the count is non-negotiable — all N runs must happen. For `--tillend`: only a verified empty round (confirmed by git diff) stops the loop. Do not end your response without invoking this skill.